*Originally posted by Fongie*

My advice is that if you're playing with a big stack, don't look at the ev line ever. It's just never going to be accurate, because variance is bigger than most people imagine - different board textures, different opponents, flop turn and river play, hand ranges (running into top or bottom of a persons hand range) and so on... Simply, not many hands are alike if they go beyond the flop. And to get an accurate sample of, for example, different river situations, you're going to have to... well, I don't know how

I am sorry but i don't understand something.

AI EV line, as long as you are not going all in at some point before the river is actually nothing more than the winning line, am i right ?

The only moment it may strive from the winning line is then when you are all in before the river.

The direction of the evline is extacly the same as the winning line only scaled by the equity, am i correct ?

I.E. with an all in where you can win say $15 or lose $10 and your chances are 30% win 68% loss 2% split (real chances while knowing all the cards):

- If you win the AI EV line would go up by 30% of the $15 you won.

- If you lose it would then go down by 68% of the $10 you lost.

If it actually works like that, the winning line should dance in a maximum range of about 5 time the square root of the number of all in hands time the size of the average pot in those hands. someone please check my maths but i think this is correct.

*Originally posted by Fongie*

If you're shortstacking however I can see that you could look at it in a large sample, because most of your play is preflop and on the flop.

What would be a descent sample size to look at ?

Could someone here nicely post a graphic with AI EV line on a large sample playing SSS (like 250k+ hands) ?